The syndrome of agrammatism is a form of language impairment that affects sentence processing and results from acquired brain damage. Treatment studies which have taken the complexity of the underlying system into account have demonstrated improvements in sentence processing. What has not been determined is how the underlying physiology of the recovering system effects or interacts with this type of intervention. Furthermore, while there is a developing body of evidence to suggest that learning environments which encourage accurate responses and prevent error responses (i.e. errorless learning) are more efficacious than traditional models of errorful learning, there has been minimal application of this learning theory towards the rehabilitation of acquired brain damage, except in the area of memory (Wilson et al., 1994). The purpose of this subproject is to apply recent advances in learning theory and cognitive linguistics to the rehabilitation of the sentence production deficits observed in this syndrome. We will test the application of errorless learning, an approach supported by parallel distributed models of learning and cognition, to the rehabilitation of sentence production deficits. We will do this by comparing the effectiveness of this approach so that of more traditional, errorful learning. Furthermore, we will compare the impact of vicariative substitutive treatment method (i.e. one designed to encourage alternative processes to assume damaged functions) that is derived from current cognitive linguistic models of sentence production, with a restitutive approach (i.e. designed to restore function) that is more typical of traditional aphasia therapy. In so doing we can address the relative limitations of the underlying physiology of the damaged system since current therapies of brain repair suggest that the potential for recovery of function may be time-dependent. By comparing errorless and errorful learning strategies and their interactions with substitutive versus restitutive treatment approaches in a single subject design, we may identify rehabilitation environments which can maximize the individual's response to rehabilitation. It is also the goal of this project to determine if improvements observed in the controlled rehabilitation environment generalize and have an effect on the individual's functional ability to communicate. Generalized improvement in sentence production would contribute to greater independence and improved quality of life. Finally, we will attempt to associate functional changes in sentence production as a result of rehabilitation with observable changes in neural activity using fMRI.